CN103246896A - Robust real-time vehicle detection and tracking method - Google Patents
Robust real-time vehicle detection and tracking method Download PDFInfo
- Publication number
- CN103246896A CN103246896A CN2013101985896A CN201310198589A CN103246896A CN 103246896 A CN103246896 A CN 103246896A CN 2013101985896 A CN2013101985896 A CN 2013101985896A CN 201310198589 A CN201310198589 A CN 201310198589A CN 103246896 A CN103246896 A CN 103246896A
- Authority
- CN
- China
- Prior art keywords
- vehicle
- tracking
- target
- predicted position
- robustness
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The invention discloses a robust real-time vehicle detection and tracking method, which mainly solves the problems that the vehicle tracking method cannot realize stable tracking, the real-time performance is poorer, the realization is more complex, and demands of people cannot be met in the prior art. The robust real-time vehicle detection and tracking method comprises the steps of collecting positive and negative samples; extracting characteristics of the positive and negative samples to obtain a cascade classifier; detecting target pictures; establishing a tracking list and predicting positions of vehicles in the tracking list by adopting an optical flow method; and conducting path tracking according to the initial positions and the predicted positions of the vehicles. By adopting the scheme, the robust real-time vehicle detection and tracking method achieves the goals of stable tracking, better real-time performance and relatively simple realization, and has very great practical values and popularization values.
Description
Technical field
The present invention relates to a kind of method for supervising, specifically, relate to a kind of robustness vehicle and detect in real time and tracking.
Background technology
Along with country the implementing gradually of policy aspect " safe city ", " the safety traffic ", the intelligent video monitoring field is rapidly developed.The used vehicle detecting system in intelligent video monitoring field relies on static background mostly at present, carry out the feature judgement and carry out trajectory predictions by extracting moving object information, yet the monitoring scene in the reality is more than theoretic complexity, all can cause very big influence to detection as the shake of camera and the conversion of background, in order to improve accuracy of detection, the normal complicated algorithm that adopts is come separating background in the prior art, though this processing mode has reached the purpose of accurate detection, can not satisfy the requirement of real-time.
Vehicle tracking is indispensable part in the supervisory system, be in recent years popular domain to the research of track algorithm, the reasonable track algorithm of effect comprises particle filter, Kalman filtering, MIL (many case-based learnings), TLD (following the tracks of-study-detect) etc. at present always.Yet because particle filter algorithm is by generating many sample points around the target, find tracking target according to the target coupling again, thereby stability is subject to the number of sample point, and sampled point is followed the tracks of unstablely very little, and sampled point is too many, and time complexity is too high.The main based target motion state of Kalman filtering algorithm model is realized following the tracks of, but the target travel in the reality is very high because of the variation randomness of scene, so the target movement model that arranges does not have changeability in advance.MIL and TLD algorithm have good robustness to tracking, and be all more reliable than other algorithm for the stability of long-time tracking, but calculated amount is too big, all is lower than 10 frames/S for the tracking of single goal, is not suitable in the middle of the actual monitoring.To this, vehicle monitoring needs to consider simultaneously that stability and the real-time of following the tracks of just can be widely used in the middle of the real life.
Summary of the invention
The object of the present invention is to provide a kind of robustness vehicle to detect in real time and tracking, mainly solve the vehicle tracking method that exists in the prior art and follow the tracks of instability, real-time is relatively poor, and realization is comparatively complicated, can not satisfy the problem of people's demand.
To achieve these goals, the technical solution used in the present invention is as follows:
A kind of robustness vehicle detects and tracking in real time, may further comprise the steps:
(1) adopts the off-line training mode from the vehicle monitoring video pictures, to intercept vehicle head or the positive sample of afterbody picture conduct, use any image that does not comprise vehicle as negative sample, and positive and negative samples is collected in the ratio of 1:2;
(2) extract the Haar-like feature of positive and negative samples, and use the adaboost algorithm to carry out off-line training to draw the cascade classifier that the judgement to vehicle strengthens step by step;
(3) input Target Photo, extract all Haar-like features of Target Photo, use cascade classifier that the Haar-like feature of Target Photo is detected identification, and judge the feature distributed intelligence of vehicle in the Target Photo according to the Haar-like feature that identifies, and when detecting vehicle the information of vehicles of record object vehicle;
(4) set up the tracking tabulation, target vehicle is added into the tracking tabulation, adopt optical flow method that the vehicle of following the tracks of in the tabulation is carried out position prediction, and judge between vehicle and predicted position whether have shelter, if there is shelter, then analyze in conjunction with movable information and the LBP texture histogram of vehicle, draw revised vehicle predicted position;
(5) carry out path trace according to initial position and the predicted position of vehicle.
In the described step (1), the size of positive samples pictures is 20 * 20 ~ 100 * 100, and the size of negative sample picture is not less than 20 * 20.
In the described step (3), information of vehicles comprises vehicle initial position message and size.
In the described step (3), carry out the preceding first initialization zone to be detected of vehicle detection, when detecting the Target Photo with vehicle, Target Photo is divided into the square network of equalization, calculate its LBP texture histogram, and with this LBP texture histogram that calculates as the coupling correction template of carrying out follow-up tracking in the step (4).
In the described step (4), the concrete computing method of vehicle predicted position are as follows when having shelter:
(4a) set that vehicle can accept by extraneous at utmost influence down under the threshold value of vehicle movement velocity variable quantity and the vehicle predicted position and the similarity threshold between the LBP texture histogram of initial target template;
(4b) the movement velocity variable quantity of two interframe vehicles before and after the calculating
And the similarity between the LBP texture histogram of the target of same size and initial target template under the predicted position in the present frame
(4c) judge the movement velocity variable quantity that calculates
And similarity
And the magnitude relationship between preset threshold of respectively controlling oneself is when the movement velocity variable quantity that calculates
And similarity
All when respectively controlling oneself preset threshold, the mode that adopts LBP to resample is revised the predicted position of vehicle, until the movement velocity variable quantity that calculates
And similarity
In at least one less than preset threshold.
Compared with prior art, the present invention has following beneficial effect:
(1) the present invention has broken through the available technology adopting particle filter, Kalman filtering, MIL (many case-based learnings), TLD traditional algorithms such as (following the tracks of-study-detect) carries out the thinking limitation of vehicle tracking, taken full advantage of the tracking stability of optical flow method, local binary pattern feature (LBP) is applied to vehicle tracking, remedied the optical flow method shortcoming that the situation of being blocked influences in tracing process, make vehicle tracking process robust more, and draw after the empirical tests, even along with the increase of following the tracks of vehicle, travelling speed of the present invention can obviously not descend yet, can satisfy the requirement of real-time follow-up fully, and reliability is higher, have outstanding substantive distinguishing features and marked improvement, be fit to large-scale promotion application.
Description of drawings
Fig. 1 is schematic flow sheet of the present invention.
Fig. 2 is the distribution schematic diagram of cascade sorter of the present invention.
Fig. 3 divides synoptic diagram for the histogrammic network of LBP texture among the present invention.
Fig. 4 is LBP resampling synoptic diagram among the present invention.
Embodiment
The invention will be further described below in conjunction with drawings and Examples, and embodiments of the present invention include but not limited to the following example.
Embodiment
In order to realize functions such as the required vehicle detection in video frequency vehicle monitoring field, tenacious tracking, direction of vehicle movement judgement simple and reliablely, as shown in Figure 1, the invention discloses a kind of robustness vehicle detects and tracking in real time, the present invention mainly comprises two parts, i.e. vehicle detection and vehicle tracking.
Wherein, vehicle detection is the prerequisite of vehicle tracking, and a kind of implementation of vehicle detection is provided in the present embodiment, and is as follows:
Adopting the off-line training mode that the vehicle in original video is carried out positive and negative samples collects, namely intercept vehicle head or the afterbody conduct positive sample of size between 20 * 20 to 100 * 100 from video pictures, the video source of positive sample can be selected from actual traffic and monitor; Use any image do not comprise vehicle as negative sample, the size minimum of negative sample can not be less than 20 * 20, and with the ratio collection of positive and negative samples by 1:2.
As shown in Figure 2, extract the multiple Haar-like feature of the positive and negative samples of having collected, use ripe adaboost algorithm to carry out off-line training and obtain the multi-stage cascade sorter, select 20 sorters in the present embodiment for use, sorter strengthens with cascade gradually to the judgement performance of vehicle, have only the judgement of having passed through cascade classifier when the Haar-like feature of test pattern panel region, could determine whether comprise vehicle in this test picture.
When carrying out vehicle detection identification, need the input Target Photo, and use the cascade classifier that has trained to carry out eigenwert and differentiate, namely calculate all Haar-like eigenwerts of Target Photo, with the Haar-like feature as target signature, in conjunction with the adaboost learning algorithm of robust, off-line learning distributes to the characteristic probability of general vehicle.For example: when implementing, from a large amount of HD videos, collect 8000 positive samples, from network, collect 16000 negative samples, the quantity of so positive sample is enough to comprise the vehicle of different automobile types, therefrom the Haar-like feature of Ti Quing can reflect the characteristic information of vehicle fully, so the discrimination to vehicle can reach 95%, and can detect minimum dimension and be 20 * 20 target vehicle, picture for complex scene and SD video still has good detection effect, reached the requirement that in the video monitoring vehicle is effectively detected, widespread use in practice.
When use is of the present invention, need first initialization zone to be detected, treat surveyed area then and carry out vehicle detection, and record detected target vehicle initial position message and size, each target is divided into impartial rectangular node, for example be divided into 8 16 * 16 grid as shown in Figure 3, calculate LBP texture histogram, and this histogram is used for the coupling correction of follow-up tracking module as the textural characteristics of To Template.
Vehicle tracking is core of the present invention, the main optical flow method of using is as main vehicle tracking algorithm among the present invention, and increase the robustness of tracking in conjunction with LBP texture histogram and vehicle movement information, make it under the long-time situation of following the tracks of and being blocked by other object, have good stability.A kind of implementation method of carrying out vehicle tracking is provided among the present invention, as follows:
At first, adopt optical flow method to carry out position prediction:
After the initialization target, obtain gray level image and the vehicle tabulation of previous frame, current frame image is obtained through after the gradation conversion
, the vehicle tracking tabulation is set up in vehicle location the unknown of present frame at this moment, target vehicle is added into follows the tracks of tabulation, adopts optical flow method that the vehicle that adds the tracking tabulation is carried out position prediction, obtains the position candidate of each tracking target in photo current
, owing to may there be the situation that vehicle is blocked by other objects of following the tracks of, this moment then
Be taken as position candidate and treat, also need movable information and the LBP texture histogram of following the tracks of vehicle are analyzed, judge whether vehicle movement is affected, preserve simultaneously
Input as next position prediction;
Secondly, carrying out state of motion of vehicle judges:
When adopting the optical flow method tracking target, if target is not blocked, then the variation of the historical speed of target all is stable, based on this conclusion, can add the motion state that householder method of the prior art just can be judged vehicle according to the velocity variations of vehicle, the present invention selects for use in conjunction with LBP texture histogram and judges the vehicle movement unusual condition, and its method is the movement velocity variable quantity of two interframe vehicles before and after calculating
And the similarity between the LBP texture histogram of the target of same size and initial target template under the predicted position in the present frame
If
With
Simultaneously greater than certain pre-set threshold, the size of this threshold value has represented that system's acceptable vehicle is by the degree of ectocine, there are abnormal conditions in the motion that shows present vehicle or are blocked by other object, need revise the predicted position of vehicle, revise the mode that adopts LBP to resample and seek the position at the vehicle place the most similar to To Template, a kind of predicted position 8 resampling modes are on every side provided in the present embodiment, as shown in Figure 4.
Afterwards, carrying out LBP revises:
In position candidate
Determine the final position of vehicle around the loca
, generate 8 sample points according to historical speed dx and dy
, wherein, dx and dy namely represent the sample point offsets amount on x axle and y direction of principal axis centered by position candidate, each point is as the candidate point of tracing positional, and the generation picture block corresponding with the To Template size
, calculate the LBP texture histogram of each picture block and obtain candidate's similarity with the LBP texture histogram calculation of To Template by same procedure
, with the most similar sample point namely
As final vehicle tracking position, return new location information
, upgrade the LBP texture histogram of To Template simultaneously.
In order to realize Continuous Observation, need judge whether also that after having calculated the latest position information of following the tracks of vehicle new vehicle occurs, namely in detection zone, continue vehicle is detected, reject according to target overlapping area information and detect and at the vehicle of following the tracks of, as establish new detected vehicle and be
If,
, think that then this detection vehicle has been in tracking mode, it is deleted from detection list, as fresh target, initialization target information and LBP texture histogram enter the circulation tracking mode with remaining vehicle.
In order to confirm the present invention's effect in actual applications, draw real-time test data table of the present invention after after tested, as shown in table 1:
According to above-described embodiment, just can realize the present invention well.
Claims (5)
1. a robustness vehicle detects and tracking in real time, it is characterized in that, may further comprise the steps:
(1) adopts the off-line training mode from the vehicle monitoring video pictures, to intercept vehicle head or the positive sample of afterbody picture conduct, use any image that does not comprise vehicle as negative sample, and positive and negative samples is collected in the ratio of 1:2;
(2) extract the Haar-like feature of positive and negative samples, and use the adaboost algorithm to carry out off-line training to draw the cascade classifier that the judgement to vehicle strengthens step by step;
(3) input Target Photo, extract all Haar-like features of Target Photo, use cascade classifier that the Haar-like feature of Target Photo is detected identification, and judge the feature distributed intelligence of vehicle in the Target Photo according to the Haar-like feature that identifies, and when detecting vehicle the information of vehicles of record object vehicle;
(4) set up the tracking tabulation, target vehicle is added into the tracking tabulation, adopt optical flow method that the vehicle of following the tracks of in the tabulation is carried out position prediction, and judge between vehicle and predicted position whether have shelter, if there is shelter, then analyze in conjunction with movable information and the LBP texture histogram of vehicle, draw revised vehicle predicted position;
(5) carry out path trace according to initial position and the predicted position of vehicle.
2. a kind of robustness vehicle according to claim 1 detects and tracking in real time, it is characterized in that, in the described step (1), the size of positive samples pictures is 20 * 20 ~ 100 * 100, and the size of negative sample picture is not less than 20 * 20.
3. a kind of robustness vehicle according to claim 1 detects and tracking in real time, it is characterized in that in the described step (3), information of vehicles comprises vehicle initial position message and size.
4. a kind of robustness vehicle according to claim 1 detects and tracking in real time, it is characterized in that, in the described step (3), carry out the preceding first initialization zone to be detected of vehicle detection, when detecting the Target Photo with vehicle, Target Photo is divided into the square network of equalization, calculate its LBP texture histogram, and with this LBP texture histogram that calculates as the coupling correction template of carrying out follow-up tracking in the step (4).
5. a kind of robustness vehicle according to claim 4 detects and tracking in real time, it is characterized in that in the described step (4), the concrete computing method of vehicle predicted position are as follows when having shelter:
(4a) set that vehicle can accept by extraneous at utmost influence down under the threshold value of vehicle movement velocity variable quantity and the vehicle predicted position and the similarity threshold between the LBP texture histogram of initial target template;
(4b) the movement velocity variable quantity of two interframe vehicles before and after the calculating
And the similarity between the LBP texture histogram of the target of same size and initial target template under the predicted position in the present frame
(4c) judge the movement velocity variable quantity that calculates
And similarity
And the magnitude relationship between preset threshold of respectively controlling oneself is when the movement velocity variable quantity that calculates
And similarity
All when respectively controlling oneself preset threshold, the mode that adopts LBP to resample is revised the predicted position of vehicle, until the movement velocity variable quantity that calculates
And similarity
In at least one less than preset threshold.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310198589.6A CN103246896B (en) | 2013-05-24 | 2013-05-24 | A kind of real-time detection and tracking method of robustness vehicle |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310198589.6A CN103246896B (en) | 2013-05-24 | 2013-05-24 | A kind of real-time detection and tracking method of robustness vehicle |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103246896A true CN103246896A (en) | 2013-08-14 |
CN103246896B CN103246896B (en) | 2016-02-10 |
Family
ID=48926409
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310198589.6A Active CN103246896B (en) | 2013-05-24 | 2013-05-24 | A kind of real-time detection and tracking method of robustness vehicle |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103246896B (en) |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104809437A (en) * | 2015-04-28 | 2015-07-29 | 无锡赛睿科技有限公司 | Real-time video based vehicle detecting and tracking method |
CN105069472A (en) * | 2015-08-03 | 2015-11-18 | 电子科技大学 | Vehicle detection method based on convolutional neural network self-adaption |
CN105138987A (en) * | 2015-08-26 | 2015-12-09 | 电子科技大学 | Vehicle detection method based on aggregation channel characteristic and motion estimation |
CN106022263A (en) * | 2016-05-19 | 2016-10-12 | 西安石油大学 | Vehicle tracking method in fusion with feature matching and optical flow method |
CN106650805A (en) * | 2016-12-13 | 2017-05-10 | 中国科学院自动化研究所 | Visual target tracking method and device |
CN106714112A (en) * | 2017-01-20 | 2017-05-24 | 西北工业大学 | Moving vehicle tracking method and system based on crowd sensing |
CN106897735A (en) * | 2017-01-19 | 2017-06-27 | 博康智能信息技术有限公司上海分公司 | The tracking and device of a kind of Fast Moving Object |
CN107122706A (en) * | 2017-03-17 | 2017-09-01 | 华南理工大学 | A kind of vehicle checking method based on growth form extreme learning machine |
CN107392210A (en) * | 2017-07-12 | 2017-11-24 | 中国科学院光电技术研究所 | A kind of target detection tracking method based on TLD algorithms |
CN107872644A (en) * | 2016-09-23 | 2018-04-03 | 亿阳信通股份有限公司 | Video frequency monitoring method and device |
CN108615365A (en) * | 2018-05-09 | 2018-10-02 | 扬州大学 | A kind of statistical method of traffic flow based on vehicle detection and tracking |
CN108710828A (en) * | 2018-04-18 | 2018-10-26 | 北京汽车集团有限公司 | The method, apparatus and storage medium and vehicle of identification object |
CN111105444A (en) * | 2019-12-31 | 2020-05-05 | 哈尔滨工程大学 | Continuous tracking method suitable for underwater robot target grabbing |
CN112433476A (en) * | 2021-01-27 | 2021-03-02 | 北京理工大学 | Robust prediction control device and robust prediction control method for networked control system of electric vehicle |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102163278A (en) * | 2011-03-03 | 2011-08-24 | 苏州市慧视通讯科技有限公司 | Illegal vehicle intruding detection method for bus lane |
CN102779267A (en) * | 2011-05-12 | 2012-11-14 | 株式会社理光 | Method and device for detection of specific object region in image |
CN102902955A (en) * | 2012-08-30 | 2013-01-30 | 中国科学技术大学 | Method and system for intelligently analyzing vehicle behaviour |
US20130073194A1 (en) * | 2011-09-15 | 2013-03-21 | Clarion Co., Ltd. | Vehicle systems, devices, and methods for recognizing external worlds |
-
2013
- 2013-05-24 CN CN201310198589.6A patent/CN103246896B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102163278A (en) * | 2011-03-03 | 2011-08-24 | 苏州市慧视通讯科技有限公司 | Illegal vehicle intruding detection method for bus lane |
CN102779267A (en) * | 2011-05-12 | 2012-11-14 | 株式会社理光 | Method and device for detection of specific object region in image |
US20130073194A1 (en) * | 2011-09-15 | 2013-03-21 | Clarion Co., Ltd. | Vehicle systems, devices, and methods for recognizing external worlds |
CN102902955A (en) * | 2012-08-30 | 2013-01-30 | 中国科学技术大学 | Method and system for intelligently analyzing vehicle behaviour |
Non-Patent Citations (1)
Title |
---|
张和颖: "基于视频的车辆检测与跟踪", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (20)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104809437B (en) * | 2015-04-28 | 2018-04-13 | 无锡赛睿科技有限公司 | A kind of moving vehicles detection and tracking method based on real-time video |
CN104809437A (en) * | 2015-04-28 | 2015-07-29 | 无锡赛睿科技有限公司 | Real-time video based vehicle detecting and tracking method |
CN105069472A (en) * | 2015-08-03 | 2015-11-18 | 电子科技大学 | Vehicle detection method based on convolutional neural network self-adaption |
CN105069472B (en) * | 2015-08-03 | 2018-07-27 | 电子科技大学 | A kind of vehicle checking method adaptive based on convolutional neural networks |
CN105138987A (en) * | 2015-08-26 | 2015-12-09 | 电子科技大学 | Vehicle detection method based on aggregation channel characteristic and motion estimation |
CN105138987B (en) * | 2015-08-26 | 2018-05-18 | 电子科技大学 | A kind of vehicle checking method based on converging channels feature and estimation |
CN106022263B (en) * | 2016-05-19 | 2019-07-09 | 西安石油大学 | A kind of wireless vehicle tracking of fusion feature matching and optical flow method |
CN106022263A (en) * | 2016-05-19 | 2016-10-12 | 西安石油大学 | Vehicle tracking method in fusion with feature matching and optical flow method |
CN107872644A (en) * | 2016-09-23 | 2018-04-03 | 亿阳信通股份有限公司 | Video frequency monitoring method and device |
CN106650805A (en) * | 2016-12-13 | 2017-05-10 | 中国科学院自动化研究所 | Visual target tracking method and device |
CN106897735A (en) * | 2017-01-19 | 2017-06-27 | 博康智能信息技术有限公司上海分公司 | The tracking and device of a kind of Fast Moving Object |
CN106714112A (en) * | 2017-01-20 | 2017-05-24 | 西北工业大学 | Moving vehicle tracking method and system based on crowd sensing |
CN107122706A (en) * | 2017-03-17 | 2017-09-01 | 华南理工大学 | A kind of vehicle checking method based on growth form extreme learning machine |
CN107392210A (en) * | 2017-07-12 | 2017-11-24 | 中国科学院光电技术研究所 | A kind of target detection tracking method based on TLD algorithms |
CN108710828A (en) * | 2018-04-18 | 2018-10-26 | 北京汽车集团有限公司 | The method, apparatus and storage medium and vehicle of identification object |
CN108615365A (en) * | 2018-05-09 | 2018-10-02 | 扬州大学 | A kind of statistical method of traffic flow based on vehicle detection and tracking |
CN108615365B (en) * | 2018-05-09 | 2021-04-13 | 扬州大学 | Vehicle flow statistical method based on vehicle detection and tracking |
CN111105444A (en) * | 2019-12-31 | 2020-05-05 | 哈尔滨工程大学 | Continuous tracking method suitable for underwater robot target grabbing |
CN111105444B (en) * | 2019-12-31 | 2023-07-25 | 哈尔滨工程大学 | Continuous tracking method suitable for grabbing underwater robot target |
CN112433476A (en) * | 2021-01-27 | 2021-03-02 | 北京理工大学 | Robust prediction control device and robust prediction control method for networked control system of electric vehicle |
Also Published As
Publication number | Publication date |
---|---|
CN103246896B (en) | 2016-02-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103246896B (en) | A kind of real-time detection and tracking method of robustness vehicle | |
Santhosh et al. | Anomaly detection in road traffic using visual surveillance: A survey | |
Zhang et al. | Deep convolutional neural networks for forest fire detection | |
CN109948582B (en) | Intelligent vehicle reverse running detection method based on tracking trajectory analysis | |
CN107123131B (en) | Moving target detection method based on deep learning | |
WO2019179024A1 (en) | Method for intelligent monitoring of airport runway, application server and computer storage medium | |
CN110084165B (en) | Intelligent identification and early warning method for abnormal events in open scene of power field based on edge calculation | |
CN105744232A (en) | Method for preventing power transmission line from being externally broken through video based on behaviour analysis technology | |
CN101799968B (en) | Detection method and device for oil well intrusion based on video image intelligent analysis | |
CN104751136A (en) | Face recognition based multi-camera video event retrospective trace method | |
Piciarelli et al. | Surveillance-oriented event detection in video streams | |
CN104966304A (en) | Kalman filtering and nonparametric background model-based multi-target detection tracking method | |
CN105426820A (en) | Multi-person abnormal behavior detection method based on security monitoring video data | |
Gong et al. | Local distinguishability aggrandizing network for human anomaly detection | |
CN111091057A (en) | Information processing method and device and computer readable storage medium | |
Su et al. | A new local-main-gradient-orientation HOG and contour differences based algorithm for object classification | |
CN101877135B (en) | Moving target detecting method based on background reconstruction | |
CN103971100A (en) | Video-based camouflage and peeping behavior detection method for automated teller machine | |
CN104077571A (en) | Method for detecting abnormal behavior of throng by adopting single-class serialization model | |
Kalva et al. | Smart Traffic Monitoring System using YOLO and Deep Learning Techniques | |
Gao et al. | Moving object detection for video surveillance based on improved ViBe | |
CN116311166A (en) | Traffic obstacle recognition method and device and electronic equipment | |
Hsieh et al. | Grid-based template matching for people counting | |
Sreekumar et al. | TPCAM: Real-time traffic pattern collection and analysis model based on deep learning | |
CN114694090A (en) | Campus abnormal behavior detection method based on improved PBAS algorithm and YOLOv5 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |